Validation of the Seattle Heart Failure Model in Japanese Heart Failure Patients

Aim: Our aim was to validate the Seattle Heart Failure Model (SHFM) in the Japanese population. Methods: The performance of SHFM was evaluated in HF patients enrolled in the West Tokyo Heart Failure (WET-HF) and National cerebral and cardiovascular center acute decompensated heart failure (NaDEF) registries. Discrimination was assessed through the c-statistic, and calibration was assessed by calibration plots among HF patients with reduced ejection fraction (HFrEF; ≦40%) and preserved ejection fraction (HFpEF; >40%). If the model had perfect calibration, the slope and intercept would equal to 1.0 and 0.0, respectively. Results: Among 2,470 patients who were survived and discharged between 2006 and 2016, 572 (23.2%) patients died during a mean follow-up of 2.1 years. Among HFrEF patients, SHFM showed good discrimination (c-statistic = 0.75) but poor calibration with overestimation of 1-year survival (slope = 0.78 and intercept = −0.22, R² = 0.97); while in HFpEF patients, SHFM showed poor discrimination (c-statistic = 0.69) with also poor calibration with underestimation of 1-year survival (slope = 1.18 and intercept = 0.16, R² = 0.91). However, after simple recalibration to the baseline survival function of Japanese patients, calibration was corrected in the only HFrEF patients (slope = 1.03 and intercept = −0.04, R² = 0.97). Conclusions: In HFrEF patients, SHFM showed good discrimination but poor calibration; but was corrected with recalibration. The SHFM did not perform adequately in HFpEF patients.